Limbic Encephalitis: The Emperor Really Has New Clothes

A 23-year-old woman presented to the emergency department with vague symptoms that began a week prior to admission. She was, however, unable to give much history. Her husband and mother reported that she was confused, had memory problems and poor concentration, and was unable to make decisions. She also appeared depressed, irritable, and anxious. She was noted to be talking to herself on more than one occasion. There was no medical or psychiatric history, no new psychosocial stressors, and no history of street-drug abuse. She was working at a desk job in a large corporation and was seen by the on-call psychiatry resident, who found her alert and responsive, but unwilling or unable to respond to questions.

She was admitted to a psychiatry ward with a differential diagnosis that included first-episode psychosis, depression with psychotic features, and anxiety NOS. She was started on olanzapine. Her confusion worsened, and she became stuporous, uncommunicative, and febrile, with a temperature of 38.7 degrees Celsius. An urgent neurological consultation was sought as her presentation at that point was now most consistent with a CNS inflammatory disease such as a meningoencephalitis.

She was started empirically on IV ceftriaxine, acyclovir, and metronidazole. However, blood and CSF chemistries and cytology were all normal, as were bacterial and viral studies. Her brain MRI was normal, and her EEG showed bilateral slowing but without epileptiform activity, a finding consistent with an organic encephalopathy. Blood and CSF samples then returned positive for anti-NMDA (N-methyl D-aspartate) receptor antibodies. She was started on IVIG and referred for gynecological evaluation to assess her pelvic content. A pelvis ultrasound followed by a CT abdomen/pelvis identified the presence of bilateral ovarian dermoid cysts. She underwent bilateral ovarian cystectomy. Pathology confirmed that the cysts were benign.

Over the course of six months, she slowly improved and was able to go back to work. Her family confirmed her near full recovery. She had no recollection of the period surrounding her initial presentation and hospitalization.

Anti-NMDA receptor encephalitis falls within the broad category of autoimmune limbic encephalitis. These syndromes affect predominantly the medial temporal lobes and orbitofrontal and frontobasal regions of the brain. The nosology of autoimmune limbic encephalitis is still evolving. One classification is based on whether the autoantibodies are directed toward cell-surface membrane or intraneuronal antigens.

Until recently, most autoimmune encephalitides were considered a manifestation of paraneoplastic syndromes. These are rare conditions due to the remote effects of tumors that lie outside the CNS. CNS injury is due to antibody-mediated neuronal damage and specifically excludes injury by tumor invasion, metastasis, or chemotherapy. The damage-inducing antibodies are directed at antigens located in the primary tumor but then cross react with neuronal molecules or related antigens to cause the neurologic disease. For example, anti-Hu antibody is associated with small-cell lung carcinoma, and anti-Ma antibody is associated with testicular carcinoma. These paraneoplastic disorders typically involve intracellular neuronal antigen targets and have cytotoxic T cell pathogenesis.

In the mid 2000s, however, a new type of limbic encephalitis with complex and often bewildering neuropsychiatric presentations was described. The seminal paper by Vitiliani (2005) identified this new form of limbic encephalitis, which was later called anti-NMDA receptor encephalitis and caused by anti-NMDA receptor antibodies. Subsequent confirmatory work was done by Dalmau and his collaborators (Dalmau 2008). In this type of limbic encephalitis, the autoantibodies are directed against neuronal surface membrane antigens and are more amenable to immunotherapy. Other identified neuronal surface antigens that can elicit damage-inducing antibodies and a limbic encephalitis include voltage-gated potassium channels, AMPA-receptors, and GABA receptors. While anti-NMDA receptor encephalitis is not considered a “conventional” paraneoplastic disorder for the above reason, it is associated with ovarian teratomas in about 60% of cases (Dalmau 2008).

Anti-NMDA receptor encephalitis typically occurs in sequential phases. The first phase begins with malaise, headache, and fatigue, followed by florid psychiatric symptoms, including anxiety, mood dysregulation, delirium, agitation, hallucinations, and delusions. In this phase, patients often present to psychiatric services, making limbic encephalitis a quintessential neuropsychiatric disorder. The second phase could be either an unresponsive or hyperactive phase. In the unresponsive phase, patients are often noncommunicative, akinetic, or catatonic; in the hyperactive phase, autonomic instability, hypertension, dyskinesias, or stereotypies predominate.

Prior to the advent of NMDA assays and an increased awareness of this condition, anti-NMDA receptor encephalitis was understandably diagnosed as first-episode psychosis or viral encephalitis. Etiological misattribution was common, since standard blood, CSF, MRI, or EEG investigations may be within normal limits or yield nonspecific abnormalities. Definitive diagnosis is made with anti-NMDA receptor antibody assays in blood and/or CSF. Once diagnosed, the treatment involves immunotherapy—intravenous immunoglobulins, high-dose corticosteroids, plasma exchange, rituximab, azathioprine, and cyclophosphamide deployed sequentially or in parallel.

At the same time, a whole-body search for a tumor is warranted as with conventional paraneoplastic disorders. If a tumor is found, resection leads to better prognosis. The recovery process is slow and may take several months. Residual cognitive-intellectual problems are common. Moreover, in most instances, patients are unable to recall the events of the prodrome and acute illness, consistent with an illness that is centered in the limbic system and disrupts hippocampal function and hence recent memory. ■

References for this article are posted at http://www.psychnews.org/update/experts_3_26.html.